Segmentation of newborn brain MRI

Segmenting neonatal brain images is challenging due to reduced MR signal-intensity contrast between tissues of interest compared to older subjects. In this paper we discuss a new strategy that we are currently employing in our neonatal brain studies and present the segmentation problem within a Bayesian framework. We show results validating this method and a previously published method from our group against a group of manually drawn segmentations. We achieve good to excellent results, on particularly challenging data, with greater speed and ease of use